1. The General Linear Model (GLM)
Klaas Enno Stephan
Wellcome Trust Centre for Neuroimaging
University College London
With many thanks to my colleagues in the FIL methods group, particularly Stefan Kiebel, for useful slides
SPM Course, ICN May 2008

2. Overview of SPM p <0.05 Statistical parametric map (SPM)
Image time-series
Kernel
Design matrix
Realignment
Smoothing
General linear model
Statistical
inference
Gaussian
field theory
Normalisation
p <0.05
Template
Parameter estimates

3. A very simple fMRI experiment
One session
Passive word listening
versus rest
7 cycles of
rest and listening
Blocks of 6 scans
with 7 sec TR
Stimulus function
Question: Is there a change in the BOLD response between listening and rest?

4. Modelling the measured data
Why?
Make inferences about effects of interest
Decompose data into effects and error
Form statistic using estimates of effects and error
How?
stimulus function
effects estimate
linear
model
statistic
data
error estimate

5. Voxel-wise time series analysis
model
specification
parameter
estimation
hypothesis
statistic
Time
Time
BOLD signal
single voxel
time series
SPM

6. Single voxel regression model
error = +
1 2 +
Time x1 x2 e
BOLD signal

7. Mass-univariate analysis: voxel-wise GLM= + y
Model is specified by
Design matrix X
Assumptions about e
N: number of scans
p: number of regressors
The design matrix embodies all available knowledge about experimentally controlled factors and potential confounds.

8. GLM assumes Gaussian “spherical” (i.i.d.) errors
sphericity = iid: error covariance is scalar multiple of identity matrix:
Cov(e) = 2I
Examples for non-sphericity:
non-identity
non-identity
non-independence

9. = + Parameter estimation y X Least squares parameter estimate
Estimate parameters such that
minimal
Least squares parameter estimate
(assuming iid error)

10. A geometric perspectivey e
Design space defined by X x2 x1

11. What are the problems of this model?
BOLD responses have a delayed and dispersed form.
HRF
The BOLD signal includes substantial amounts of low- frequency noise.
The data are serially correlated (temporally autocorrelated)  this violates the assumptions of the noise model in the GLM

12. Problem 1: Shape of BOLD response Solution: Convolution model
hemodynamic response function (HRF)
The response of a linear time-invariant (LTI) system is the convolution of the input with the system's response to an impulse (delta function).
expected BOLD response
= input function impulse response function (HRF)

13. Convolution model of the BOLD response
Convolve stimulus function with a canonical hemodynamic response function (HRF):
 HRF

14. Problem 2: Low-frequency noise Solution: High pass filtering
S = residual forming matrix of DCT set
discrete cosine transform (DCT) set

15. High pass filtering: example
blue = data
black = mean + low-frequency drift
green = predicted response, taking into account low-frequency drift
red = predicted response, NOT taking into account low-frequency drift

16. Problem 3: Serial correlations
with
1st order autoregressive process: AR(1)
autocovariance
function

17. Dealing with serial correlations
Pre-colouring: impose some known autocorrelation structure on the data (filtering with matrix W) and use Satterthwaite correction for df’s.
Pre-whitening: Use an enhanced noise model with hyperparameters for multiple error covariance components Use estimated autocorrelation to specify filter matrix W for whitening the data.

18. How do we define W? Enhanced noise model
Remember how Gaussians are transformed linearly
Choose W such that error covariance becomes spherical
Conclusion: W is a function of V  so how do we estimate V?

19. Multiple covariance components
enhanced noise model V Q1 Q2 = 1
+ 2
Estimation of hyperparameters  with ReML (restricted maximum likelihood).

20. Contrasts & statistical parametric maps
Q: activation during listening ?
Null hypothesis:

21. t-statistic based on ML estimates
ReML-estimate

22. Physiological confounds
head movements
arterial pulsations
breathing
eye blinks
adaptation affects, fatigue, fluctuations in concentration, etc.

23. Correlated and orthogonal regressorsy x2
x2* x1
Correlated regressors =
explained variance is shared between regressors
When x2 is orthogonalized with regard to x1, only the parameter estimate for x1 changes, not that for x2!

24. Outlook: further challenges
correction for multiple comparisons
variability in the HRF across voxels
slice timing
limitations of frequentist statistics  Bayesian analyses
GLM ignores interactions among voxels  models of effective connectivity

25. Correction for multiple comparisons
Mass-univariate approach: We apply the GLM to each of a huge number of voxels (usually > 100,000).
Threshold of p<0.05  more than 5000 voxels significant by chance!
Massive problem with multiple comparisons!
Solution: Gaussian random field theory

26. Variability in the HRF
HRF varies substantially across voxels and subjects
For example, latency can differ by ± 1 second
Solution: use multiple basis functions
See talk on event-related fMRI

27. Summary Mass-univariate approach: same GLM for each voxel
GLM includes all known experimental effects and confounds
Convolution with a canonical HRF
High-pass filtering to account for low-frequency drifts
Estimation of multiple variance components (e.g. to account for serial correlations)
Parametric statistics